Short-let revenue management lives or dies on two things: how much comparable supply exists in a neighbourhood, and what that supply actually costs a guest once cleaning and fees are added. Neither is a field you can read; both have to be assembled.
The source is named because it is public; the client is not. Every figure on this page is a measured production number the client agreed to publish. Have a source of your own? Send it over, whatever its size — you get an answer within 24 hours.
calendar-nights since 2023
a year
measured coverage
listings tracked
Not more rows. Every one of these projects started with somebody who already had data and could not use it for the decision in front of them.
The problem
The client prices short-let inventory for property managers. Their competitive picture came from a sample of listings collected once a week, which meant a manager pricing a weekend in a busy neighbourhood was doing it against a picture assembled before the neighbourhood filled up. What they needed was the whole comparable set, with calendars, refreshed daily.
Fixed on day one
Every one of these is measured continuously and reported on the same dashboard the client watches. A commitment nobody measures is a sentence in a proposal.
None of these is solved by better headers or a bigger proxy pool. Each needs a different piece of engineering, and working out which one you are actually facing is most of the job.
Comparability is local — three streets over is a different market — so a city-level sample tells a manager almost nothing about their own block.
What we doMarkets are collected by geographic tiling sized to observed price variance, dense where prices change over short distances and sparse where they do not.
A listing's value is in its calendar, and the calendar is not one number. Collecting a year of availability for millions of listings is the expensive half of the problem.
What we doCalendar depth is tiered by how far ahead the client's managers actually price: the near window is refreshed daily, the far window weekly, and the split is measured against how often far dates change.
On a two-night stay a cleaning fee can be a third of the total. A nightly rate without it is not comparable to anything.
What we doLanded price is computed at the party sizes and stay lengths the client models, with every fee component stored separately so managers can see where the money is.
Short-let supply turns over fast, and a static listing set decays into fiction within a quarter.
What we doSupply discovery runs continuously alongside refresh, so new listings enter the picture within a day and delistings are recorded as events rather than as silence.
The architecture is not interesting; every extraction system has a queue, a fetcher and a parser. These are the decisions that made this one work where the last one did not.
Tile density comes from measured price variance across distance, not from a fixed grid. Dense neighbourhoods get dense tiling; a quiet suburb does not, which is where the cost saving lives.
The near booking window is refreshed daily and the far window weekly, with the boundary set by how often far dates actually move. Collecting 365 days uniformly is the naive plan that makes this unaffordable.
Fees are captured as components and the landed total is computed at the client's modelled party sizes. A nightly rate on its own is the number most competitors publish and nobody can use.
New supply is found by ongoing traversal rather than by a periodic rebuild of the listing set, so a neighbourhood filling up is visible within a day.
Adding a market must not add cost per market. Tiling, tiering and discovery are all tuned against a per-market budget that is reported weekly.
Production figures, not a benchmark run. Coverage is recounted daily against an independent sample of the live source rather than asserted, which is why the numbers are not round.
Daily profile
One calendar request returns about sixty nights, so 3.4 million requests — 39 a second — produce 210 million night-values a day: 6.3 billion a month, 77 billion a year, 230 billion since 2023. The near window is refreshed daily and the far window weekly; collecting all 365 nights for all 7.4 million listings every day would need nine times the traffic and change nothing a revenue manager acts on.
Headline
These four sit in the support agreement. When one of them drifts outside its band, we are alerted within fifteen minutes and fixing it is routine work under the monthly arrangement, not a change request.
The columns are the client's own numbers from before the rebuild and the measured ones from production today. The left column is the part most vendors would rather not put on a page.
| Metric | Before | Today |
|---|---|---|
Comparable set |
A weekly sample | The whole neighbourhood, daily |
Price basis |
Nightly rate | Landed total at modelled party sizes |
Calendar visibility |
None | 365 days, tiered refresh |
Cost of a new market |
Linear | Flat, budget-driven |
Managers stopped arguing with the tool the month the landed price appeared. A nightly rate that nobody pays is the fastest way to lose a user's trust.
One case file per source, each opening the same way: what the client needed, what the source put in the way, how the pipeline is built and what the numbers are today.
46M itineraries priced a day, decomposed
Read the case →
94k routes, 210M fare quotes a day
Read the case →
340M listings watched, sold prices captured
Read the case →
84M listings, eight markets, eight currencies
Read the case →
64M offers, competition at the offer level
Read the case →
Start here
Send the source, the fields you need and roughly how often. You get a straight answer within 24 hours: whether it can be done, what makes it hard, what coverage is achievable and roughly what it costs to build and to run. Whatever the size.
Direct
NDA before technical detail, as always. If it isn't our kind of work, we say so in the first reply.